J W
@polymath-wannabe.bsky.social
Engineer, technologist, scientist, communicator
Cybersecurity, weather, and infectious disease junkie
Builder of things. Interested in too many topics for my own good.
* Skeets not representative of my employer *
Cybersecurity, weather, and infectious disease junkie
Builder of things. Interested in too many topics for my own good.
* Skeets not representative of my employer *
All that said, the discussion has been a learning experience for me, both on the topic and with respect to BlueSky culture. As with all such things, it seems, the cost was being added to a couple of block lists.
Life is tradeoffs.
Life is tradeoffs.
November 21, 2025 at 2:45 PM
All that said, the discussion has been a learning experience for me, both on the topic and with respect to BlueSky culture. As with all such things, it seems, the cost was being added to a couple of block lists.
Life is tradeoffs.
Life is tradeoffs.
...that even the most creative person is, in fact, building on what others have done.
Regardless, this is an issue we will be addressing far into the future.
Regardless, this is an issue we will be addressing far into the future.
November 21, 2025 at 1:47 PM
...that even the most creative person is, in fact, building on what others have done.
Regardless, this is an issue we will be addressing far into the future.
Regardless, this is an issue we will be addressing far into the future.
In science, there is a tension between the free use and availability of knowledge, and credit or even lockdown of intellectual property. In creative arts, from what I've seen, things lean toward emphasizing the latter, even though a strong argument can be made...
November 21, 2025 at 1:47 PM
In science, there is a tension between the free use and availability of knowledge, and credit or even lockdown of intellectual property. In creative arts, from what I've seen, things lean toward emphasizing the latter, even though a strong argument can be made...
The above-mentioned symbolic math programs are an example. If I get a result from Mathematica, it generally doesn't cite the mathematicians who came up with the methodology being used to solve a given equation, for example.
November 21, 2025 at 1:47 PM
The above-mentioned symbolic math programs are an example. If I get a result from Mathematica, it generally doesn't cite the mathematicians who came up with the methodology being used to solve a given equation, for example.
None of this is intended to wade into ethical issues with respect to how the tools are actually being applied. The better ones link to their sources, at least.
But other tools exist that harvest or provide time saving application of knowledge, including for commercial use.
But other tools exist that harvest or provide time saving application of knowledge, including for commercial use.
November 21, 2025 at 1:47 PM
None of this is intended to wade into ethical issues with respect to how the tools are actually being applied. The better ones link to their sources, at least.
But other tools exist that harvest or provide time saving application of knowledge, including for commercial use.
But other tools exist that harvest or provide time saving application of knowledge, including for commercial use.
In doing so, I admit I didn't realize the degree to which the entire topic is a minefield.
November 21, 2025 at 1:25 PM
In doing so, I admit I didn't realize the degree to which the entire topic is a minefield.
potentially adding to mental health problems for certain individuals. What I posted originally was intended to address the first point, not delve into a comprehensive philosophical discussion of the use of AI.
November 21, 2025 at 1:25 PM
potentially adding to mental health problems for certain individuals. What I posted originally was intended to address the first point, not delve into a comprehensive philosophical discussion of the use of AI.
The original point of the thread revolved around how the tool is both adding to problems professionals experience with being bombarded by nonsense (also seen all over social media in pretty much any scientific discussion) and...
November 21, 2025 at 1:25 PM
The original point of the thread revolved around how the tool is both adding to problems professionals experience with being bombarded by nonsense (also seen all over social media in pretty much any scientific discussion) and...
I personally don't think the issue of ethics is entirely cut and dry when it comes to knowledge use (by humans or otherwise). But I get that there is strong feeling on the issue, and for good reason.
My intent here was really not to wade into a debate of all aspects of the ethics of AI.
My intent here was really not to wade into a debate of all aspects of the ethics of AI.
cited.one
November 21, 2025 at 1:25 PM
I personally don't think the issue of ethics is entirely cut and dry when it comes to knowledge use (by humans or otherwise). But I get that there is strong feeling on the issue, and for good reason.
My intent here was really not to wade into a debate of all aspects of the ethics of AI.
My intent here was really not to wade into a debate of all aspects of the ethics of AI.
Thank you for coming back to the conversation, and I apologize if my username is distracting. I am, in fact, a degreed electronics engineer and astrophysicist.
Though this is hardly my area of expertise, I do understand at least some of the objections, which seem to be more creative than technical.
Though this is hardly my area of expertise, I do understand at least some of the objections, which seem to be more creative than technical.
November 21, 2025 at 1:25 PM
Thank you for coming back to the conversation, and I apologize if my username is distracting. I am, in fact, a degreed electronics engineer and astrophysicist.
Though this is hardly my area of expertise, I do understand at least some of the objections, which seem to be more creative than technical.
Though this is hardly my area of expertise, I do understand at least some of the objections, which seem to be more creative than technical.
At this point, I see LLMs as knowledge-gathering and synthesis tools. They are imperfect (to say the least),
November 21, 2025 at 1:07 PM
At this point, I see LLMs as knowledge-gathering and synthesis tools. They are imperfect (to say the least),
As I've stated elsewhere: it's not really that different from using a symbolic math program. The solutions presented are likely right. But being able to understand and work with calculus allows a double-check.
November 21, 2025 at 1:07 PM
As I've stated elsewhere: it's not really that different from using a symbolic math program. The solutions presented are likely right. But being able to understand and work with calculus allows a double-check.
And that is perhaps the most pernicious aspect: if you don't have the expertise in the first place to double check results, what an LLM gives you (generally stated with confidence) may very well be garbage. How can you tell?
November 21, 2025 at 1:07 PM
And that is perhaps the most pernicious aspect: if you don't have the expertise in the first place to double check results, what an LLM gives you (generally stated with confidence) may very well be garbage. How can you tell?
I'm a degreed scientist and engineer. My honest opinion at this point is that LLMs have their uses, especially when it comes to time savings. But you'd better not treat the output as a black box - the work has to be checked, and available to be checked, including calculations and methodology.
November 21, 2025 at 1:07 PM
I'm a degreed scientist and engineer. My honest opinion at this point is that LLMs have their uses, especially when it comes to time savings. But you'd better not treat the output as a black box - the work has to be checked, and available to be checked, including calculations and methodology.
But a fair amount seems to go farther. One can actually state that one finds the use of LLMs in many cases to be problematic, and it's apparently not enough.
November 21, 2025 at 1:07 PM
But a fair amount seems to go farther. One can actually state that one finds the use of LLMs in many cases to be problematic, and it's apparently not enough.
Thank you, I suppose.
November 21, 2025 at 12:49 PM
Thank you, I suppose.
And just a note: the LLM actually provided the methodology, which for the most part simply employed astropy and related libraries.
If you already have Python code written for any situation, fantastic.
If you already have Python code written for any situation, fantastic.
November 21, 2025 at 12:44 PM
And just a note: the LLM actually provided the methodology, which for the most part simply employed astropy and related libraries.
If you already have Python code written for any situation, fantastic.
If you already have Python code written for any situation, fantastic.
Yes, there are. Which is what was used for comparison.
I find the hostility with respect to the use of a new tool remarkable.
I find the hostility with respect to the use of a new tool remarkable.
November 21, 2025 at 12:41 PM
Yes, there are. Which is what was used for comparison.
I find the hostility with respect to the use of a new tool remarkable.
I find the hostility with respect to the use of a new tool remarkable.
Well, that depends. In the first instance, it ended up saving me quite a bit of time off of coding from scratch. In the second (spectral line fitting), I verified the answers manually via other methodology, as it was a test. The system (Gemini) did very well.
November 21, 2025 at 7:56 AM
Well, that depends. In the first instance, it ended up saving me quite a bit of time off of coding from scratch. In the second (spectral line fitting), I verified the answers manually via other methodology, as it was a test. The system (Gemini) did very well.
Garbage in, garbage out remains pretty applicable.
November 21, 2025 at 7:29 AM
Garbage in, garbage out remains pretty applicable.
To me, it's not all that different than a symbolic math program: super useful in certain circumstances, but having the solution to a complex equation in hand doesn't do much for you if you got the equation wrong in the first place. Etc.
November 21, 2025 at 7:29 AM
To me, it's not all that different than a symbolic math program: super useful in certain circumstances, but having the solution to a complex equation in hand doesn't do much for you if you got the equation wrong in the first place. Etc.
I suppose I should clarify that I care little for what some of the AI advocates/evangelists push. A lot of it is hype at best and ridiculous or even dangerous at worst, and I think there will be big retrenchments.
I've been taking things on a case by case basis
I've been taking things on a case by case basis
November 21, 2025 at 7:29 AM
I suppose I should clarify that I care little for what some of the AI advocates/evangelists push. A lot of it is hype at best and ridiculous or even dangerous at worst, and I think there will be big retrenchments.
I've been taking things on a case by case basis
I've been taking things on a case by case basis